Security for buildings and other locations often involves the use of mounted video devices, such as but not limited to video cameras. Such video devices, which may be fixed or movable, obtain a series of images of one or more scenes. These images are processed, either manually (e.g., where a human monitor reviews the obtained images) and/or at least partially automatically by image processors (e.g., computers or other processing devices) to analyze the obtained images according to particular algorithms and to catalogue and/or act on the result. When automated, intelligent image processing is used at least in part, such processing can be made more efficient and consistent.
One example use of mounted video devices with at least partial automatic image processing is task-based intelligent motion detection (IMD). IMD methods process incoming images provided by the mounted video devices to determine whether sufficient motion is present in certain locations within a scene. The sensitivity, or threshold determined amount of change between images to determine that motion has occurred, typically can be selected for individual locations within a scene. As a nonlimiting example, one or more locations within a scene can be selected (e.g., marked as a sensitive area) to detect motion. This is useful to mask out areas within a scene having inherent motion (as just one example, trees).
Within IMD generally, several types of motion detection are possible. Nonlimiting examples of IMD functionality include loitering persons detection, removed objects detection, idle objects detection, objects within range detection, objects moving against the flow detection, and tamper detection. For example, with loitering persons detection, an image processor may be configured to detect whether a person remains within a scene for a particular amount of time.
Current IMD techniques are provided generally in two settings. One conventional IMD setting is in the form of software residing on a computer (e.g., PC) linked to a video device via a network. The computer, executing the software, processes the video received from the mounted video devices.
A second setting for IMD is an embedded solution within a fixed video device, wherein one or more processors within the fixed video device itself are configured for processing images using one or more types of IMD functionality. By embedding the processors within the fixed video device itself, the fixed video device can view a scene and produce a series of images, process the images according to IMD, and even take certain actions without the requirement of being on a network. Such integrated IMD solutions also allow video devices to provide a modular security solution by being incorporated into a network and passing along video and results of IMD for further processing and/or action.
Movable video devices, such as mounted video cameras, on the other hand, currently present problems for image processing and object detection. One example movable mounted camera, a PTZ (pan-tilt-zoom) camera, moves in 3D space. The three dimensions of the PTZ camera are defined by pan, tilt, and zoom, respectively. A set of pan, tilt, and zoom positions defines an overall position.
The present inventors have recognized that the users of such movable video devices also have a need for intelligent motion detection techniques such as (but not limited to) loitering persons detection, object removal detection, etc., which reside in the camera itself, to provide benefits such as (but not limited to) those provided by incorporating IMD image processing in a fixed camera. However, currently no solution to such need exists.